Dicyandiamide Modification Extends Durability of Iron-Based Fuel Cell Catalysts

Iron-based catalysts designed to replace platinum in hydrogen fuel cells often fail within hours of operation, but a new open-access study reports a dicyandiamide modification strategy that dramatically extends their working lifetime. The research, published in the journal Catal, describes an atomically dispersed Fe-N-C electrocatalyst that ran for 264 hours under constant voltage in a hydrogen-air proton exchange membrane fuel cell and lost only 23% of its peak power density after 30,000 voltage cycles in accelerated durability testing. The work addresses what the authors identify as a critical bottleneck for commercializing platinum-group-metal-free fuel cells.

Fe-N-C catalysts are attractive because their acidic oxygen reduction activity rivals that of platinum-carbon benchmarks in rotating disk electrode tests, thanks to distinctive electronic and geometric structures at iron sites coordinated by nitrogen. In real membrane electrode assemblies, however, these materials decay severely, typically losing about half of their initial performance within 15 hours. The team attributes this decay primarily to two mechanisms: demetallation, in which iron atoms leach out of the FeN4 active sites, and corrosion of the surrounding carbon support under the harsh oxidative conditions inside an operating cell.

Earlier spectroscopic work cited in the study distinguished two dominant FeN4 configurations in Fe-N-C materials using iron-57 Mössbauer spectroscopy. A high-spin S1 moiety with pyrrolic coordination, designated FeN4C12, is highly active but unstable, while a low- or intermediate-spin S2 moiety with pyridine coordination, FeN4C10, offers superior stability at some cost in activity. The enhanced stability of the pyridinic sites stems from a stronger iron-nitrogen bond arising from their geometry, which makes the iron center less vulnerable to oxidative attack and acid leaching. Prior studies have shown that enriching these stable sites, for example through chemical vapor deposition, can extend catalyst lifetime.

The new synthesis builds on this insight with a two-step procedure. Zeolitic imidazolate framework ZIF-8 particles, roughly 200 nanometers across, were first pyrolyzed under argon to form a nitrogen-doped carbon support. After acid etching and impregnation with an iron(III) salt, the material was ground with dicyandiamide at a mass ratio of 1:2 and pyrolyzed a second time. During this step, dicyandiamide thermally condenses into nitrogen-containing intermediates that coat the carbon surface, then decomposes into pyridinic nitrogen sites that capture isolated iron atoms, forming pyridinic FeN4 configurations. A reference catalyst was prepared identically without dicyandiamide.

Multiple characterization techniques confirmed the atomic dispersion of iron in both materials. X-ray diffraction showed no iron crystalline phases, aberration-corrected scanning transmission electron microscopy revealed isolated bright spots corresponding to single iron atoms, and inductively coupled plasma mass spectrometry measured iron contents of about 5.22 weight percent for the modified catalyst and 5.14 percent for the reference. X-ray absorption near-edge structure and extended X-ray absorption fine structure analyses placed the average iron oxidation state between zero and plus three, dominated by Fe3+, and found no iron-iron scattering path near 2.20 angstroms that would indicate metallic aggregates. Wavelet transform analysis further distinguished the atomically dispersed Fe-N4 moieties from metallic iron or iron oxide nanoparticles.

The key structural differences emerged from Mössbauer spectroscopy and Raman measurements. The modified catalyst contained 32% of the stable pyridinic D2 moiety with a D2/D1 ratio of 0.60, compared with 20% and 0.30 for the unmodified material, demonstrating that dicyandiamide drives iron sites toward the more stable pyridinic configuration during pyrolysis. The modified catalyst also showed enhanced graphitization: a lower ratio of disordered-to-graphitic carbon bands in Raman spectra, intensified carbon-carbon double bond features near 285.9 electron volts in carbon K-edge XANES, and short-range graphitic domains visible in high-resolution electron microscopy. Its nitrogen 1s X-ray photoelectron spectrum showed a pyridinic-to-pyrrolic nitrogen ratio of 1.73 versus 1.49 for the reference.

In rotating disk electrode tests in oxygen-saturated 0.1 M perchloric acid, the modified catalyst achieved a half-wave potential of 0.85 volts, slightly below the 0.86 volts of the unmodified catalyst, confirming the expected trade-off between stability and intrinsic activity. Its electron transfer number approached 4.0, indicating a dominant four-electron oxygen reduction pathway, and rotating ring-disk measurements recorded a hydrogen peroxide yield of just 2.02% between 0.2 and 0.8 volts, lower than the 3.25% and 2.40% measured for the reference catalyst and commercial platinum-carbon respectively. Reduced peroxide formation matters because hydrogen peroxide generates free radicals that attack both the iron sites and the carbon scaffold.

Durability results were striking. After 30,000 cycles between 0.6 and 0.95 volts, the modified catalyst’s half-wave potential dropped by only 23 millivolts, comparable to commercial platinum-carbon at 25 millivolts, while the unmodified catalyst lost 64 millivolts. Electrochemical stripping of nitrite quantified active site density, showing a loss of only 6.1% after 10,000 cycles versus 57.5% for the reference. Under a harsher 0.95 to 1.3 volt cycling regime probing carbon corrosion, the modified catalyst lost just 8 and 16 millivolts after 5,000 and 25,000 cycles, while the reference lost 40 millivolts after only 5,000. In chronoamperometric testing at 0.6 volts, the modified catalyst retained 80% of its initial current after 500 hours, whereas the reference retained only 57% after 107 hours.

In full membrane electrode assemblies under hydrogen-oxygen conditions at 100 kilopascals, the modified catalyst delivered 50.4 milliamperes per square centimeter at 0.9 volts, exceeding the US Department of Energy 2025 target of 44.0, along with 1.05 amperes per square centimeter at 0.60 volts and a peak power density of 1.0 watts per square centimeter. That peak power was below the 1.17 and 1.36 watts per square centimeter of the reference and platinum catalysts, but stability diverged sharply under DOE accelerated stress protocols: the modified assembly lost 8% peak power after 5,000 cycles and 23% after both 20,000 and 30,000 cycles, with only 60 millivolts of voltage loss at 0.8 amperes per square centimeter, against 228 millivolts for the reference, which decayed progressively to nearly complete inactivation within 50 hours of constant-voltage operation. Energy-dispersive mapping showed the modified catalyst retained more iron after cycling, dropping from 1.60 to 0.81 weight percent versus 1.95 to 0.41 for the reference.

Density functional theory calculations explained the mechanism. Pyridinic FeN4 sites showed demetallation energy barriers of 3.823 electron volts, and defect-containing pyridinic sites 2.792 electron volts, far above the 0.652 and 1.265 electron volts of the corresponding pyrrolic structures, indicating much greater resistance to iron leaching. Charge density analysis revealed stronger electron transfer between pyridinic sites and the carbon substrate, and free-energy diagrams showed the pyridinic configuration suppresses the hydrogen peroxide-forming two-electron pathway. The authors conclude that dicyandiamide simultaneously stabilizes the active sites, reinforces the graphitic carbon support, and curbs peroxide generation, offering a general route toward robust platinum-group-metal-free electrocatalysts, though the reported performance figures remain laboratory-scale measurements rather than validated commercial deployments.

Subject of Research: Dicyandiamide modification of Fe-N-C electrocatalysts to improve durability in proton exchange membrane fuel cells

Article Title: Boosting the durability of Fe-N-C electrocatalysts for PEM fuel cells by dicyandiamide modification strategy

Article References: Lin, X., Lu, Z., Zhao, L., Han, M., Wang, Y., Huang, S., Ling, H., Huang, Y., Yun, J., & Cao, D. (2026). Boosting the durability of Fe-N-C electrocatalysts for PEM fuel cells by dicyandiamide modification strategy. Catal, 2(1), Article 12. https://doi.org/10.1007/s44422-026-00023-z

Image Credits: AI Generated

DOI: 10.1007/s44422-026-00023-z

Keywords: Fe-N-C electrocatalyst, PEM fuel cells, oxygen reduction reaction, dicyandiamide, pyridinic FeN4 sites, durability, PGM-free catalyst, Mössbauer spectroscopy, carbon corrosion, hydrogen peroxide, DFT calculations, ZIF-8